xtensor-stack/xtensor-python

Python bindings for xtensor

What it solves

It bridges the gap between high-performance C++ array manipulation and the Python ecosystem. It allows developers to use NumPy arrays directly within C++ code without copying data, enabling the use of C++'s speed and advanced features while maintaining compatibility with Python's data structures.

How it works

It uses the pybind11 library to create seamless bindings between C++ and Python. It provides two primary C++ containers—pyarray and pytensor—that wrap NumPy arrays in-place. These containers expose the xtensor library's features, such as broadcasting and lazy evaluation, directly to the NumPy data.

Who it’s for

This library is designed for developers and library authors who need to write high-performance C++ extensions for Python, specifically for applications involving multi-dimensional arrays and numerical computing.

Highlights

  • Enables in-place manipulation of NumPy arrays using C++.
  • Supports NumPy-style broadcasting and lazy computing.
  • Provides pyarray for dynamic dimensions and pytensor for faster, compile-time dimension handling.
  • Includes a pyvectorize tool to easily turn C++ scalar functions into vectorized functions callable from Python.
  • Header-only library with seamless integration via the conda package manager.

Related

  • Project
    songshikang0111/awesome-REA-cases

    Awesome REA Cases is an unofficial, actively maintained collection of 817 reverse-engineering projects, experiments, and ideas where AI tools like Claude and REA help revive old devices, rebuild apps, preserve games, and understand hidden behavior. Each case links to public evidence and notes limits, updated six times daily.

  • Project
    intent-hq/intent

    Intent is a desktop app that coordinates AI coding agents, notes, tasks, git and terminals via a Rust backend daemon (`intentd`). All state lives locally in SQLite, accessed through JSON‑RPC over a Unix socket or secure WebSocket. The packaged app bundles the daemon; advanced users can install the daemon separately. It supports multiple AI providers, optional GitHub/Linear/Sentry integrations, and an upcoming iOS companion. No telemetry is sent; updates are pulled from GitHub releases only.

  • Project
    ben-vargas/ai-sdk-provider-claude-code

    A community-maintained provider that enables the use of the Claude Agent SDK and Claude Code CLI within the Vercel AI SDK ecosystem, supporting advanced agentic features like tool use and structured output.

  • Project
    exsinger-hub/Sivia

    Sivia is a Codex/Claude Code plugin that converts a paper’s method description into an editable PowerPoint figure, using a curated knowledge base of scientific‑figure examples and supporting iterative revisions.